—There are many researcher homepages on Web, if one wants to process researcher information for search engine, building a semantic profile for the academic researcher to identify and annotate information is an effective method. In this paper, we label Chinese researcher information with Conditional Random Fields (CRF) model, which has achieved good performance on Named Entity Identification. We proposed a hybrid annotation method which combines Conditional Random Fields and semantic rules, considering some features such as suffix, prefix, and semantic features of named entity at the same time. The comparison experiments show that this method can correctly extract the real content of the Chinese researcher homepages and assign a suitable category label to each part of the contents simultaneously.

Chinese Researcher Profile Annotation Based on Conditional Random Fields with Semantic Rules

As the world continues to embark on its era of innovation, recent technological advancements easily become integrated into people's everyday living. With the establishing premise of artificial intelligence (AI) on the horizon, technology becomes increasingly transparent to its users (termed as ambient technology) in an attempt to mimic human nature when handling complex processes and delivering efficient results. In spite of this, majority of research on voice user interfaces (VUIs) have been partial to the benefit of people with disabilities. This study overwhelms this limitation as the implementation of a VUI in a smart classroom is achieved for the purposes of ambient technology. The main objective of this study is to develop a far-field speech-controlled smart classroom with natural language processing (NLP) built under KNX standard for appliance control. Using NLP techniques such as shallow parsing and Naive Bayes classification, the researchers were able to write their own API in Python for a VUI codenamed Luna. From the experiment conducted in this study, Luna was able to obtain a confusion matrix (CM) accuracy of 95.56% with the implementation of the Naive Bayes classifier alongside the shallow parser.

Far-field Speech-controlled Smart Classroom with Natural Language Processing built under KNX Standard for Appliance Control

To this date, most of the Chinese Semantic Role Labeling (SRL) methods based on deep learning still rely on complex feature or architecture engineering, which more or less weaken the generalization ability of model and increase the complexity of this task. In this article, we proposed a deep hybrid model, which is composed of Bi-directional Long Short Term Memory (Bi-LSTM), Self-Attention and Conditional Random Field (CRF), as an end to-end system without using any pre-defined features. We just took the sentences that have been segmented as input, and depended on the competence of model to extract potential context dependency information. After that, we used sentence level tag information via a CRF layer to get the final labeling results. Experimental results on Chinese Proposition Bank (CPB) shows that our model is really effective for this task, which can achieve 81.2% F1 score.

Chinese Semantic Role Labeling with Hybrid Model

In recent years, the amount of digital data that we produce has increased exponentially. This flood of information, often referred to as “big data,”	 is creating both opportunities and challenges in all areas of life. In the domain of biology, technology has enabled us to sequence the genomes of humans and many other organisms, but we are far from understanding the biological roles played by all of these genes. The Gene Ontology seeks to address this problem by annotating genes to terms describing biological processes, molecular functions, and cellular components. However, the ontology’s manual curators cannot keep up with the rate at which information is being discovered and published. Hence, there is a need for computational methods that can rapidly process the biomedical literature and suggest new annotations for verification. This study uses support vector machines to predict Gene Ontology annotations for Saccharomyces cerevisiae (yeast). I tested the usefulness of two types of literature features: co-occurrence of gene names in articles, and co-occurrence in abstracts of gene names with keywords taken from GO term definitions. My results demonstrate that support vector machines using literature co-occurrence data as features can predict GO annotations with high accuracy. In many cases where simple gene-gene co-occurrence does not work well, better results can be obtained using gene-keyword co-occurrence. I found that a very simple text mining strategy —	 identifying words that occur in only one GO term definition —	 was an effective way of choosing keywords. Although predictions based on gene-gene co-occurrence and those based on gene-keyword co-occurrence were highly correlated, there are terms for which one set of predictions was significantly more accurate than the other. I was able to combine the two sets of predictions effectively using a voting scheme in which gene-gene predictions were weighted at 70% and gene-keyword predictions at 30%.

Predicting Gene Ontology Annotations Based on Literature Co-Occurrence

There is a widely held belief in the NLP and computational linguistics communities that identifying and defining roles of predicate arguments in a sentence has a lot of potential for and is a significant step toward improving important applications such as document retrieval, machine translation, question answering and information extraction. In this paper, we present an semantic role labeling (SRL) system for Chinese that exploits many aspects of the rich features of the languages. Finally, we compare system based on CRFs and SVMs. The experiment yields a global SRL FB1 score of 92.89%.

Chinese semantic role labeling using CRFs and SVMs

We present two different shallow parsing methods that use language independent features The first system does not split shallow parsing into any sub-tasks, and it is called one-phase shallow parsing system The second one is called two-phase system that decomposes shallow parsing into two main sub-tasks, recognition (chunk) and classification (chunk), which are performed sequentially and independently with separate modules Both systems are machine learning based systems, making the use of SVM classifiers The main focus of this paper is to investigate the impact of different methods to the performance of shallow parsing Experimental results show that one-phase method is quite effective; in fact it yields better performance between the two methods Although each individual model was quite strong to begin with, we found that the two-phase method actually degraded the performance, or brought zero improvement Although more sophisticated linguistic features are helpful, we use the same features in two shallow parsing methods

Chinese shallow parsing using different phase systems

To be able to represent the whole hierarchical phrase structure, 10 types of Chinese chunks are defined. The paper presents a method of Chinese shallow Paring based on Support Vector machines (SVMs). Conventional recognition techniques based on Machine Learning have difficulty in selecting useful features as well as finding appropriate combination of selected features. SVMs can automatically focus on useful features and robustly handle a large feature set to develop models that maximize their generalizability. On the other hand, it is well known that SVMs achieve high generalization of very high dimensional feature space. Furthermore, by introducing the Kernel principle, SVMs can carry out the training in high-dimensional space with smaller computational cost independent of their dimensionality. The experiments produced promising results.

Yongmei Tan

Papers

Chinese semantic role labeling using CRFs and SVMs

Chinese shallow parsing using different phase systems

Applying Support Vector Machines to Chinese Shallow Parsing